In order to accurately detect the faults of electrical automation equipment, this paper proposes a data-based neural network to diagnose the faults of electrical automation equipment. This study investigates various (Radio Access Network, RAN) architectures such as cloud-RAN, heterogeneous cloud-RAN, and fog-RAN. These architectures are examined in various contexts, including system efficiency, spectrum and energy efficiency, fronthaul capacity, latency, resource sharing and allocation, etc. A neural network structure based on the (back propagation) model is used to perform forward computation on the sampled raw data, error Calculations and errors are backpropagated. On this basis, the self-adaptive learning fault detection algorithm is used to realize the self-adaptive fault detection of automatic electrical equipment. In addition to being able to accurately determine the known state of the device, the algorithm is also able to self-study the state of a non-training sample set, allowing the detection of device failures through adaptation. The experimental results show that the method is reliable, the error detection rate is greater than 0.95, and it has good anti-noise performance.
Cite this article as: Z. You, Q. Wu, H. Zhou, J. Ming and R. Boddu, "Fault detection of electrical automation remote equipment based on data network," Electrica, 23(3), 458-465, 2023.